Prediction and Analysis of Meteorological Drought Based on Time Series(Case Study: SALMAS Watershed)

Time series analyses is a base method for more decisions about hydrological process and water operation. In Iran, drought is a continues and normal condition happening frequently and can be predicted by statistical and mathematical methods and models. In this research, meteorological drought was analyzed in 5 stations of Salmas region (Nazarabad, Salmas, Yalgozagaj, Chahrygeolya and Urban) located in West Azarbijan Province, Iran. For this purpose, the monthly rainfall time series, 30 years lengths (1979-2008) for these stations were simulated by times series models as: AR, MA, ARMA and ARIMA. Then the best model was used for forecasting the drought. Then the monthly rainfalls were predicted for next 7 years using the chosen model. In the following, the intensity of droughts was calculated by standard precipitation index (SPI). The final model chosen for the stations of Nazarabad, Salmas, Yalgozagaj, Chahrygeolya and Urban were the ARMA(1,1), ARMA(3,2), ARMA(1,1), ARMA(2,3) and ARMA(4,2) models, respectively. The results showed that for 30 years (1978-2008), drought had increasing trend just in Nazarabad station, while in other stations such as Salmas, Yalgozagaj, Chahrygeolya had decreasing trend.